石油价格预测算法的仿真研究

摘要

Study the oil price change trend problem. Oil price changes are of nonlinear, variability and randomicity, so the traditional forecasting methods based on linear modeling cannot describe the changing rule of oil prices,and thus prediction accuracy is relatively low. In order to improve the oil price prediction accuracy and an oil price prediction model is put forward based on wavelet neural network. Oil price changing trend is forecasted based on wavelet analysis' excellent analysis function and neural network's nonlinear predictive power. The simulation results show that wavelet neural network has improved the forecasting accuracy, reduced the predicting error compared with the oil price prediction model, and can reflect the change rule of oil prices. It is an effective, high precision oil price predicting tools.%研究石油价格变化趋势问题,由于价格的不确定性,石油价格变化具有非线性、突变性和随机性.传统预测方法均是基于线性建模的,无法很好描述石油价格变化规律,从而预测精度较低.为提高石油价格预测精度,提出一种小波神经网络的石油价格预测模型.根据小波分析优异的多尺变分析功能和神经网络非线性预测能力对石油价格变化趋势进行仿真预测.仿真结果表明,相对于传统石油价格预测模型,小波神经网络提高了预测精度,降低了预测误差,更好的反映了石油价格的变化规律,是一种有效、高精度的石油价格预测工具,为预测石油的价格提供了参考.